U.S. patent application number 10/121846 was filed with the patent office on 2002-12-05 for electrically controllable integrated optical cross-connect.
Invention is credited to Wei, Haiqing, Xue, Xin.
Application Number | 20020181855 10/121846 |
Document ID | / |
Family ID | 26819881 |
Filed Date | 2002-12-05 |
United States Patent
Application |
20020181855 |
Kind Code |
A1 |
Xue, Xin ; et al. |
December 5, 2002 |
Electrically controllable integrated optical cross-connect
Abstract
An integrated optical cross-connect device and associated
methods are described, the cross-connect device comprising a
plurality M of input waveguides formed in a first material layer of
an integrated circuit, a plurality N of output waveguides formed in
a second material layer of the integrated circuit, and a plurality
MN of micromechanically actuated bridge elements formed in at least
one intermediate layer lying between the first and second material
layers. Responsive to an electrical control signal, each bridge
element establishes an index-guided, nonreflecting optical path
between its associated input waveguide and its associated output
waveguide. Preferably, the bridge element comprises an arcuate
waveguide structure substantially surrounded by air or other
nonsolid material, the arcuate waveguide structure being twistably
connected to a remainder of the intermediate layer by a narrow neck
portion. When electrostatically actuated, one end of the arcuate
waveguide structure rises to meet the input waveguide while the
other end lowers to meet the output waveguide. Associated
fabrication methods and an expandable, modular cross-connect system
based on the cross-connect device are also described.
Inventors: |
Xue, Xin; (Sunnyvale,
CA) ; Wei, Haiqing; (Sunnyvale, CA) |
Correspondence
Address: |
Ivan S. Kavrukov
Cooper & Dunham LLP
1185 Avenue of the Americas
New York
NY
10036
US
|
Family ID: |
26819881 |
Appl. No.: |
10/121846 |
Filed: |
April 12, 2002 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
60283568 |
Apr 13, 2001 |
|
|
|
Current U.S.
Class: |
385/23 ;
385/17 |
Current CPC
Class: |
G02B 6/3584 20130101;
G02B 6/12002 20130101; G02B 6/3556 20130101; H04Q 11/0005 20130101;
G02B 6/356 20130101; G02B 6/3508 20130101; G02B 6/3536 20130101;
G02B 2006/12145 20130101; H04Q 2011/003 20130101; G02B 6/357
20130101 |
Class at
Publication: |
385/23 ;
385/17 |
International
Class: |
G02B 006/35 |
Claims
What is claimed is:
1. An optical switch, comprising: a first waveguide having an input
for receiving a light beam, and an output; a second waveguide
nonintersecting with said first waveguide and having an output; and
a bridge waveguide movable between a first position and a second
position, said bridge waveguide in said first position forming an
index-guided, substantially nonreflecting optical pathway between
said first and second waveguides such that said light beam exits
said second waveguide, said bridge waveguide in said second
position not providing said pathway so that said light beam exits
said first waveguide.
2. The optical switch of claim 1, said optical switch being formed
in a planar integrated circuit comprising a vertical stack of
substantially parallel material layers, said first waveguide being
formed in first of said material layers, said second waveguide
being formed in a second of said material layers, and said bridge
waveguide being formed in a third of said material layers.
3. The optical switch of claim 2, wherein said third material layer
is positioned between said first and second material layers.
4. The optical switch of claim 3, said first and second waveguides
being nonparallel, said first waveguide passing over said second
waveguide at a first angle around a common vertical axis, said
bridge waveguide comprising an input end substantially parallel to
said first waveguide and separated therefrom by a first gap when
said bridge waveguide is in said second position, said bridge
waveguide comprising an output end substantially parallel to said
second waveguide and separated therefrom by a second gap when said
bridge waveguide is in said second position, said input end of said
bridge element being urged across said first gap toward said first
waveguide when said bridge waveguide is in said first position such
that a first optical contact therewith is achieved, said output end
of said bridge element being urged across said second gap toward
said second waveguide when said bridge waveguide is in said second
position such that a second optical contact therewith is achieved,
said bridge waveguide bending the light beam by said first angle
while guiding it between said input end and said output end.
5. The optical switch of claim 4, said input end of said bridge
element physically contacting said first waveguide when said bridge
waveguide is in said first position.
6. The optical switch of claim 4, said input end of said bridge
waveguide remaining physically separate from said first waveguide
when said bridge waveguide is in said first position, but being
positioned sufficiently close thereto such that a resonant optical
coupling is formed therewith.
7. The optical switch of claim 4, wherein said first angle is less
than 90 degrees.
8. The optical switch of claim 7, wherein said first angle is less
than 60 degrees.
9. The optical switch of claim 8, said bridge waveguide being
arcuate in shape and having a radius of curvature at all locations
therealong greater than a bending loss threshold.
11. The optical switch of claim 4, said bridge waveguide comprising
an electrostatically active material, said input and output ends
being urged toward said first and second waveguides, respectively,
by electrostatic forces.
12. The optical switch of claim 4, said bridge waveguide comprising
an arcuate tongue-like structure, said input end and output end
being on opposite sides of said tongue-like structure, said
tongue-like structure being substantially surrounded by
motion-permitting gaps and being connected to a remainder of the
third material layer by a flexible, narrow neck that permits
vertical movement of said input and output ends.
13. The optical switch of claim 4, wherein said bridge waveguide is
piezoelectrically, magnetostrictively, or photorestrictively
actuated.
14. An optical cross-connect device formed in an integrated
circuit, comprising: a plurality of input waveguides formed in a
first layer of the integrated circuit; a plurality of output
waveguides formed in a second layer of the integrated circuit, said
first and second layers being separated by at least one
intermediate layer, each output waveguide passing underneath each
input waveguide at a distinct cross-connect location; and at each
cross-connect location, a micromechanically actuated bridge element
formed in said at least one intermediate layer for optically
coupling its associated input waveguide to its associated output
waveguide responsive to an electrical control signal, said bridge
element establishing an index-guided optical path between the input
waveguide and the output waveguide when in an ON position, said
bridge element not establishing said optical path when in an OFF
position.
15. The optical cross-connect of claim 14, said input waveguides
being substantially parallel to each other, said output waveguides
being substantially parallel to each other and oriented at a first
angle with respect to said input waveguides, each of said bridge
elements comprising an input end that optically couples to the
associated input waveguide for receiving a light beam traveling
therein, an output end that optically couples to the associated
output waveguide for transferring said light beam thereto, and an
arcuate index-guiding waveguide that incrementally bends the light
beam by said first angle as it guides the light beam from said
input end to said output end.
16. The optical cross-connect of claim 15, wherein said first angle
is less than 90 degrees, and wherein said arcuate waveguide has a
radius of curvature greater than a bending loss threshold at all
points therealong.
17. The optical cross-connect of claim 16, wherein said first angle
is less than 60 degrees.
18. The optical cross-connect of claim 15, wherein said input end
of said bridge element is moved toward said input waveguide until
mechanical contact therewith is established for achieving said
optical coupling.
19. The optical cross-connect of claim 15, wherein said input end
of said bridge element is moved toward said input waveguide by an
amount that does not establish mechanical contact but that brings
said input end into close enough proximity with said input
waveguide to establish a resonant coupling therewith.
20. The optical cross-connect of claim 15, wherein said bridge
element is piezoelectrically, magnetostrictively, or
photorestrictively actuated.
21. The optical cross-connect of claim 15, wherein said bridge
element is electrostatically actuated.
22. The optical cross-connect of claim 15, said bridge element
being substantially surrounded by gaps in said at least one layer
of the integrated circuit such that vertical motion of the input
and output ends thereof are permitted, said arcuate index-guiding
waveguide being connected to a remainder of said at least one layer
by a neck portion that permits twisting of the arcuate
index-guiding waveguide around an axis substantially parallel to
said first and second layers such that said input end rises toward
said first layer when said output end lowers toward said second
layer.
23. An expandable, modular optical cross-connect system,
comprising: at least one M.times.N integrated optical cross-connect
module, M>1, N>1, each cross-connect module comprising: a top
surface, a bottom surface, a first side opposite a second side, and
a third side opposite a fourth side; a plurality M of input
waveguides extending from said first side to said second side, said
first and second sides being mateably shaped to each other such
that two such cross-connect modules placed side-by-side in a
direction of said input waveguides achieve flush optical
communication with each other, light beams exiting input waveguides
on the second side of one cross-connect module entering
corresponding input waveguides on the first side of the other
cross-connect module; a plurality N of output waveguides extending
from said third side to said fourth side, said third and fourth
sides being mateably shaped to each other such that two such
cross-connect modules placed side-by-side in a direction of said
output waveguides achieve flush optical communication with each
other, light beams exiting output waveguides on the fourth side of
one cross-connect module entering corresponding output waveguides
on the third side of the other cross-connect module; a plurality MN
of controllable cross-connecting elements that optically couple any
of said M input waveguides to any of said N output waveguides
responsive to at least one externally provided electrical control
signal; and at least one electrical contact located on said bottom
surface for receiving said at least one externally provided
electrical control signal; a jM.times.kN optical switching
backplane capable of holding a j.times.k array of said
cross-connect modules, j>1, k.ltoreq.1, comprising: an array
surface having a j.times.k array of electrical contacts positioned
to establish electrical communication with said electrical contacts
of said cross-connect modules when placed on said array surface; an
input connector comprising j adjacent sets of M optical input
elements, each set being mateably shaped to achieve flush optical
communication with said first side of said cross-connect module;
and an output connector comprising k adjacent sets of N optical
output elements, each set being mateably shaped to achieve flush
optical communication with said fourth side of said cross-connect
module; whereby said optical cross-connect system is expandable
from an M.times.N device when a single cross-connect module is
inserted into said optical switching backplane to an aM.times.bM
device when provided with (ab-1) additional cross-connect modules,
1<a.ltoreq.j, 1.ltoreq.b.ltoreq.k.
24. The expandable, modular optical cross-connect system of claim
23, each of said cross-connect modules having exactly four
sides.
25. The expandable, modular optical cross-connect system of claim
24, each of said four sides being flat and having surface
variations of no more than 0.4 .mu.m in areas corresponding to said
input and output waveguides, whereby an air gap of no more than 0.8
.mu.m exists between any two respective input waveguides or any two
respective output waveguides for any two adjacent cross-connect
modules.
26. The expandable, modular optical cross-connect system of claim
25, said input waveguides of each cross-connect module being formed
in a first material layer thereof, said output waveguides of each
cross-connect module being formed in a second layer thereof, each
controllable cross-connecting element comprising a
micromechanically actuated bridge waveguide establishing an
index-guided optical path between one of the input waveguides and
one of the output waveguides when in an ON position.
27. The expandable, modular optical cross-connect system of claim
26, wherein said bridge waveguide is piezoelectrically,
magnetostrictively, or photorestrictively actuated responsive to
said externally provided electrical control signals.
28. The expandable, modular optical cross-connect system of claim
26, wherein said bridge waveguide is electrostatically actuated
responsive to said externally provided electrical control
signals.
29. A method for fabricating an integrated optical cross-connect
comprising an upper waveguide in an upper layer, a lower waveguide
in a lower layer, and a flexible bridging element in an
intermediate layer therebetween, the flexible bridging element
being patterned in said intermediate layer and being at least
partially suspended therefrom into a subsurface air gap,
comprising: forming said lower waveguide layer; forming a lower
spacer layer above said lower waveguide layer using a first
material in a first lateral area corresponding to said subsurface
air gap and using a second material in a second lateral area
corresponding to a base of said flexible bridging element, said
first material being etchable by a first etchant to which said
second material is etch-resistant; forming said intermediate layer
above said lower spacer layer using said second material in a third
lateral area corresponding to said flexible bridging element
including said base and using said first material in a fourth
lateral area corresponding to said subsurface air gap, said fourth
lateral area being vertically coincident with said first lateral
area in at least one lateral location; forming an upper spacer
layer above said intermediate layer using said first material in a
fifth lateral area corresponding to said subsurface air gap and
using a third material in a sixth area corresponding to said base
of said flexible bridging element, said third material also being
etch-resistant to said first etchant, said fifth lateral area being
vertically coincident with said fourth lateral area in at least one
lateral location; forming said upper layer above said upper spacer
layer using a fourth material, said fourth material being etchable
by a second etchant to which said first material is etch-resistant;
masking off said upper layer at locations corresponding to the
upper waveguide and said base of said flexible bridging element;
etching non-masked areas of said upper layer until said upper
spacer layer is exposed in said non-masked areas; and isotropically
etching said first material from said upper spacer layer, said
intermediate layer, and said lower spacer layer using said first
etchant, said subsurface air gap being formed around said flexible
bridging element.
30. A method for fabricating an integrated optical cross-connect
comprising an upper waveguide in an upper layer, a lower waveguide
in a lower layer, and a flexible bridging element in an
intermediate layer therebetween, the flexible bridging element
being patterned in said intermediate layer and being at least
partially suspended therefrom into a subsurface air gap,
comprising: on a first substrate, forming said lower waveguide
layer using a first material; forming a lower spacer layer above
said lower waveguide layer using a second material, said second
material being etchable by a first etchant to which said first
material is etch-resistant; forming said intermediate layer above
said lower spacer layer using a third material that is also
etch-resistant to said first etchant; masking off said intermediate
layer at locations corresponding to the flexible bridging element;
etching non-masked areas of said intermediate layer until said
lower spacer layer is exposed in said non-masked areas;
isotropically etching away said first material from said lower
spacer layer until said intermediate layer is undercut beneath said
flexible bridge element by an mount sufficient to allow a desired
amount of flexibility thereof; on a second substrate, forming said
upper waveguide layer; forming an upper spacer layer above said
upper waveguide layer; masking off said upper spacer layer at
locations corresponding to a base of the flexible bridging element;
etching non-masked areas of said upper spacer layer until said
upper waveguide layer is exposed in said non-masked areas; and
wafer-bonding said upper spacer layer on said second substrate to
said intermediate layer on said first substrate.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application claims the benefit of Provisional
Application Ser. No 60/283,568, filed Apr. 13, 2001, which is
incorporated by reference herein.
FIELD
[0002] This patent specification relates to optical devices. More
particularly, it relates to optical cross-connect devices for use
in optical networks or in other applications.
BACKGROUND
[0003] Optical cross-connect switches may be used in a variety of
practical applications including optical communication
applications. For example, in a fiber optic communications network
in which each fiber optic cable carries a plurality of
wavelength-division multiplexed (WDM) channels, it may be desirable
to switch traffic from a source fiber to any one of a plurality of
destination fibers, without requiring the demultiplexing or
electrical-to-optical conversion of the optical channels. Such
fiber-based switching may be desirable, for example, to allow
traffic switching around cable cuts, to better balance
communications traffic among communications centers, or for a
variety of other purposes in either long-haul or short-haul
networks. It would be desirable to provide an optical cross-connect
device capable of directing an optical signal from any of "M" input
optical waveguides to any of "N" optical output waveguides,
responsive to electrical control signals. It would be further
desirable to provide such an optical cross-connect device in which
M and N may be relatively large, while the overall device size
remains relatively modest. It would be still further desirable to
provide an optical cross-connect device that is readily amenable to
known semiconductor manufacturing methods, for allowing both
smaller size and lower per-unit costs. It would be even further
desirable to provide an optical cross-connect device that is
readily amenable to a single-growth fabrication process hat avoids
the need for multiple growth and wafer bonding steps, thereby
further increasing fabrication yields and reducing costs. It would
be still further desirable to provide an optical cross-connect
device that is amenable to inclusion in a modular, expandable
optical cross-connect system.
SUMMARY
[0004] An integrated optical cross-connect device is provided,
comprising a plurality M of input waveguides formed in a first
material layer of an integrated circuit, a plurality N of output
waveguides formed in a second material layer of the integrated
circuit, and a plurality MN of micromechanically actuated bridge
elements formed in at least one intermediate material layer lying
between the first and second material layers. Responsive to an
electrical control signal, each bridge element establishes an
index-guided, nonreflecting optical path between its associated
input waveguide and its associated output waveguide when in an ON
position. In an OFF position, the bridge waveguide optically
isolates that input waveguide from that output waveguide.
[0005] In one preferred embodiment, each input waveguide passes
near each output waveguide at a common vertical axis, the input and
output waveguides forming a first angle with respect to each other.
A light beam traveling along the input waveguide is bent by an
amount corresponding to that first angle when it is transferred
over to the output waveguide by the bridge waveguide. Preferably,
the first angle is less than 60 degrees to reduce bending losses,
and the bridge waveguide has an arcuate shape having a radius of
curvature greater than a bending loss threshold at all locations
therealong. In one preferred embodiment, in the ON position, the
bridge waveguide makes mechanical contact with the input and output
waveguides to establish the optical path therebetween. In another
preferred embodiment, in the ON position, the bridge waveguide does
not establish mechanical contact with the input and output
waveguides, but is positioned close enough thereto to establish
resonant couplings therewith to establish the optical path.
[0006] Preferably, the bridge waveguide is substantially surrounded
by air gaps or other non-solid material such that vertical motion
of its ends is permitted. The bridge element is connected to a
remainder of the intermediate layer by a twistable neck portion.
The neck portion twists when one end of the bridge waveguide rises
to meet the input waveguide while the other end lowers to meet the
output waveguide. In one preferred embodiment, he bridge element
comprises an electrostatically active material and is actuated by
electrostatic forces. In other preferred embodiments, the bridge
element is piezoelectrically, magnetostrictively, or
photorestrictively actuated. The bridge element, which is suspended
in a subsurface air gap, is fabricated by constructing the
requisite device patterns in an entirely solid multi-layer format,
with a uniquely etchable material being disposed in all areas in
which there is going to be an air gap. The other material used to
form the waveguides and the bridge element, as well as a bridge
element base, are selected to be etch-resistant to the etchant of
the uniquely etchable material. During subsequent device
fabrication steps, the uniquely etchable material is etched away,
leaving the requisite structure including the suspended bridge
element.
[0007] In other preferred embodiments, an expandable, modular
optical cross-connect system is provided comprising a backplane for
holding a j.times.k array of the M.times.N cross-connect modules
and providing optical and electrical signals thereto. The input and
output waveguides of each cross-connect module run across its
entire length and width, respectively, to precisely flattened side
edges thereof. When two such cross-connect modules are positioned
next to each other in the lengthwise direction, light beams exit
the input waveguides of one module and proceed across a very narrow
gap directly into corresponding input waveguides of the next module
with tolerable losses. Similar optical connectivity is achieved
among modules positioned next to each other in the widthwise
direction, and therefore an aM x bN cross-connect device can be
formed by inserting ab modules on the backplane,
1.ltoreq.a.ltoreq.j, 1.ltoreq.b.ltoreq.k.
BRIEF DESCRIPTION OF THE DRAWINGS
[0008] FIG. 1 illustrates an exterior view of an optical
cross-connect device in accordance with a preferred embodiment;
[0009] FIG. 2 illustrates a top cut-away view of a portion of an
optical cross-connect device in accordance with a preferred
embodiment;
[0010] FIG. 3 illustrates a simplified perspective view of a
waveguide bridging portion of the optical cross-connect device of
FIG. 2 in an "OFF" state;
[0011] FIG. 4 illustrates a simplified perspective view of a
waveguide bridging portion of he optical cross-connect device of
FIG. 2 in an "ON" state;
[0012] FIG. 5 illustrates a simplified side cross-section of a
portion of the optical cross-connect device of FIG. 2;
[0013] FIGS. 6-9 illustrate steps for fabricating two lower layers
and a waveguide bridging layer in accordance with a preferred
embodiment;
[0014] FIGS. 10-11 illustrate steps for fabricating two upper
layers and wafer bonding the resulting structure to the structure
of FIG. 9 to form an optical cross-connect device in accordance
with a preferred embodiment;
[0015] FIGS. 12-15 illustrate steps for fabricating an optical
cross-connect device in a ingle-growth process in accordance with a
preferred embodiment; and
[0016] FIGS. 16-17 illustrate a modular, expandable optical
cross-connect system in accordance with a preferred embodiment.
DETAILED DESCRIPTION
[0017] FIG. 1 illustrates an electrically controllable integrated
optical cross-connect device 100 in accordance with a preferred
embodiment. The objective of the cross-connect device 100 is to
selectively and dynamically connect any of a first plurality of
planar optical waveguide inputs IN(1), IN(2), . . . , IN(M) to any
of a second plurality of planar optical waveguide outputs OUT(1),
OUT(2), . . . , OUT(N) responsive to electrical control signals.
Because the design is readily amenable to construction using known
semiconductor fabrication methods, the size M.times.N of the
cross-connect device may grow very large, e.g., up to
100.times.100. At the same time, however, the overall size of the
cross-connect device 100 remains relatively modest, e.g., 2
cm.times.2 cm, making it useful for a variety of practical
applications. The electrical signals are shown as entering the
cross-connect device 100 through pins on its side. In an
alternative preferred embodiment described further infra, these
electrical signals may enter through surface-mounted metallic
connections on the top and/or bottom of the cross-connect device
100.
[0018] FIG. 2 illustrates a top view of a portion of the layout of
the cross-connect device 100, in particular showing four (4)
waveguide bridging elements 202, 204, 206, and 208 at the
respective intersections of two input waveguides IN(x) and IN(x+1)
with two output waveguides OUT(y) and OUT(y+1). The remainder of
this disclosure focuses on a single waveguide bridging element 202,
it being understood that there are (MN-1) other waveguide bridging
elements that are substantially similar to the waveguide bridging
element 202. The waveguide bridging element 202 represents one type
of bridging waveguide element that is separate from the input and
output waveguides themselves, and that is used to bridge the input
and output waveguides responsive to electrical control signals.
[0019] According to a preferred embodiment, the waveguide bridging
element 202 is a movable waveguide element designed and configured
such that it couples light from the input waveguide IN(x) into the
output waveguide OUT(y) when in an "ON" state. When switched into
the "ON" state, the waveguide bridging element 202 physically moves
into a position such it provides an optical pathway connecting the
waveguides. When in an "OFF" state, the waveguide bridging element
202 physically moves to an "OFF" position such that the optical
circuit between the input waveguide IN(x) into the output waveguide
OUT(y) is cut off. The movement of the waveguide bridging element
202 between the "ON" and "OFF" positions is actuated by the
electrical control signals provided to the optical cross-connect
device 100 shown in FIG. 1. While the embodiments described infra
relate to an electrostatically actuated movement, it is to be
understood that the scope of the preferred embodiments is not so
limited, and that actuation may be provided using piezoelectric,
magnetostrictive, photorestrictive, or other microelectromechanical
(MEMS) actuation techniques in accordance with the preferred
embodiments.
[0020] For clarity of description, the remainder of the present
disclosure presents the input waveguide IN(x) as being in a
"bottom" layer (although shown as solid lines in FIG. 2), the
output waveguide OUT(x) as being in a "top" layer (although shown
as dotted lines in FIG. 2), and the waveguide bridging element 202
as being in a "middle" layer. An upper spacer layer is placed
between the top and middle layers, and a lower spacer layer is
placed between the middle and bottom layers. With reference to FIG.
2, the input and output waveguides IN(x) and OUT(y) are shown at an
angle .theta. with respect to each other around a common vertical
axis, this angle being close to 90 degrees. However, it has been
found that use of a higher angle, e.g. at 120 degrees or above,
provides for better light coupling between the input and output
waveguide arrays. In the case of .theta.=120 degrees, for example,
the coupling is superior to the case of .theta.=90 degrees because
the light does not need to change direction by the full 90 degrees,
but rather only needs to change direction by 60 degrees, thereby
reducing bending losses in the waveguide bridging element 202.
[0021] Waveguide bridging element 202 comprises a thin (e.g., 10-20
.mu.m) layer of material shaped to form an anchor portion 210, a
neck portion 212, an electrostatically active portion 214, and an
arc connector portion 216. It is to be appreciated that the
waveguide bridging element 202, while presented infra as a single
layer, may actually comprise multiple layers over its entire area,
or over selected portions of its area, to achieve the described
functionalities. Importantly, however, the selected materials must
have common etching characteristics with respect to a common set of
etchants to allow the device to be properly fabricated using the
methods described infra. The material used for waveguide bridging
element 202 in its arc connector portion 216 should be a dielectric
material having refractive index that is appreciably large as
compared to the immediately surrounding void regions. This allows
for incoming light to be turned by the angle (180-.theta.) from the
input waveguide IN(x) to the output waveguide OUT(y). Examples of
such materials may include quartz (crystalline SiO.sub.2) or
polycrystalline silicon.
[0022] The material for the electrostatically active portion 214
requires at least a portion of its thickness to comprise a
conducting material such as a metal or a doped semiconductor. This
allows the electrostatically active portion 214 to experience an
induced charge responsive to the presence of a voltage differential
between two external plates (described further infra), resulting in
a torque around the neck portion 212. The electrostatically active
element 214 shares an elongated border with the arc connector
portion 216, as illustrated in FIG. 2, such that these two elements
will move together.
[0023] In an alternative preferred embodiment, the
electrostatically active element 214 is completely integrated with
the arc connector portion 216 from a lateral perspective, i.e., the
electrostatically active element 214 simply lies along a top or
bottom surface of the arc connector portion 216. In this
alternative preferred embodiment, there will be some optical losses
due to the presence of a conducting material along a surface of a
waveguide. It has been found, however, that a highly conductive
material is not necessarily required to achieve the needed
electrostatic activity. Instead, only a partially conducting
material can be used, such as a more lightly doped semiconductor,
for cutting down on the optical losses. Indeed, it has been found
that even an end-to-end resistance of 1 megohm for the
electrostatically active element 214 provides sufficient
electrostatic properties to achieve the required torque around the
neck portion 212. In another alternative preferred embodiment,
electrostatically active element 214 and the arc connector portion
216 can actually be the same physical body, e.g., a doped
semiconductor. For simplicity and clarity of explanation, the
alternative embodiment in which the electrostatically active
element 214 lies along a top surface of the arc connector portion
216 is presented herein.
[0024] The electrostatically active element 214 is hingeably
connected to the anchor portion 210 via the neck portion 212, such
that the arc connector portion 216 will twist about an axis of
rotation, the axis of rotation being parallel to the planes of the
input and output waveguides, the axis of rotation being oriented at
an angle that bisects the angle .theta. formed between the IN(x)
and OUT(y) waveguides. Neck portion 212 and anchor portion 210
comprise a material that is solid, but that is flexible enough to
allow the neck portion 212 to rotate by a very small angle
responsive to a torque applied to the arc connector portion 216.
Although any of a variety of materials may satisfy this
requirement, quartz or polycrystalline silicon are two exemplary
candidates for these elements. Quartz or single-crystal silicon may
be particularly advantageous because, due to their crystalline
structure, it is robust against material fatigue that may occur
near the neck portion. In many cases, all of the portions of the
waveguide bridging element 202 may share a common layer of material
such as quartz or crystalline silicon.
[0025] By way of nonlimiting example, typical lateral dimensions of
the optical interconnect device of FIG. 2 may include a waveguide
spacing "a" of about 0.1 mm to 1 mm, and a waveguide width "w" of
about 1-10 .mu.m. Another typical lateral dimension may include a
dimension "b" of the arc connector portion 216 of about 0.1 mm to
0.5 mm.
[0026] FIG. 3 illustrates a simplified perspective view of the arc
connector portion 216 of the waveguide bridging element 202 in an
"OFF" state. As indicated in FIG. 3, the input waveguide IN(x) is
placed in a first plane, while the output waveguide OUT(y) is
placed in a second plane distinct from, but parallel to, the first
plane. The input waveguide IN(x) passes under the output waveguide
OUT(y) at the common vertical axis 350. The waveguide bridging
element 202 containing the arc connector portion 216 is positioned
in a third plane lying between the first and second planes. In the
embodiment of FIG. 3, the arc connector portion 216 comprises the
electrostatically active element 214 on its surface. As indicated
in FIG. 3, the arc connector portion 216 contains as shallow a
bending radius as is practicable to avoid bending losses,
preferably having a bending radius therealong not less than a
minimum threshold bending radius. For clarity of disclosure, the
substrates into which the waveguides IN(x) and OUT(y) are formed
are not shown in FIG. 3. Also for clarity of disclosure, none of
the neck portion 212, the anchor portion 210, or spacer layers is
shown in FIG. 3.
[0027] Plates 302 and 304 are positioned as shown in FIG. 3 and are
switchably connected to a voltage source V. The plates 302 and 304
are usually integrated into the same substrates as the waveguides
IN(x) and OUT(y), and may be either metallic or may comprise a
heavily doped semiconductor material. Where no voltage exists
between the plates 302 and 304, the waveguide bridging element is
in an "OFF" state. As indicated
[0028] FIG. 4 illustrates a simplified perspective view of the arc
connector portion 216 of the waveguide bridging element 202 in an
"ON" state. Where a sufficient voltage V exists between the plates
302 and 304, there is an (a) induced charge distribution along the
electrostatically active element 214, (b) an electrostatic
attraction between the respective ends of the arc connector portion
216 and the plates 302 and 304, (c) a resulting torque at the neck
portion 212, and (d) a resulting movement of the arc connector
portion 216 into the position shown in FIG. 4. As illustrated in
FIG. 4, the arc connector portion 216 now touches each of the
waveguides IN(x) and OUT(y). It has been found that this touching
is sufficient to establish an optical circuit between the two, with
acceptably low signal loss.
[0029] In another preferred embodiment, the arc connector portion
216 is positioned close to, but not touching, the waveguides IN(x)
and OUT(y) in the "ON" state. In particular, the waveguide bridging
element 202 is dimensioned and positioned such that, when in the
"ON" state, the two ends of the arc connector portion 216 form
resonant couplings with the waveguides IN(x) and OUT(y),
respectively. Precise dimensions, waveguide elasticity
characteristics, and electrostatic force characteristics are
required in order to establish proper spacing to achieve the
resonant coupling. For this reason, the preferred embodiment in
which the arc connector portion 216 makes actual contact with the
waveguides is generally easier to manufacture and implement.
Nevertheless, the preferred embodiment in which resonant coupling
is achieved may be useful in achieving variable coupling
efficiency, so as to allow dynamic control of the output optical
signal strength. This may be useful, for example, in achieving a
one-to-many "broadcast" capability described further infra.
[0030] In another preferred embodiment, electrostatic force on the
arc connector portion 216 may be achieved even without the
existence of electrostatically active portion 214 and even if arc
connector portion 216 is purely dielectric. This is possible if a
tip electrode or a linear electrode is configured (e.g., by making
it a pointed electrode or a narrow-banded electrode) to create a
non-uniform electric field along at least a portion of the
dielectric arc connector portion 216. The electrostatic force will
arise from induced electric dipoles in the dielectric arc connector
portion 216. The induced dipoles will experience a nonzero force
due to the non-uniform electric field. In this embodiment, however,
the strength of the neck portion 212 would need to be substantially
weaker as compared to the two-plate embodiment, because the induced
torque will be substantially lower.
[0031] FIG. 5 illustrates a conceptual side cut-away view of the
optical cross-connect device 100 taken along a slice K-K'
illustrated in FIG. 2. Formed in a lower substrate 502 is both the
input waveguide IN(x) and the bottom electrode plate 302. Between
the waveguide bridging element 202 and the lower substrate 502 is a
lower spacer layer 506. Formed in an upper substrate 504 is both
the output waveguide OUT(y) and the top electrode plate 304, these
elements being illustrated in dotted lines because, as shown in
FIG. 2, they lie outside of the plane K-K', these elements being
included in FIG. 5 for completeness and clarity of description.
Between the waveguide bridging element 202 and the upper substrate
504 is an upper spacer layer 508.
[0032] By way of example and not by way of limitation, exemplary
dimensions for the cross-connect device as illustrated in FIG. 5
may include: a thickness for lower substrate layer 502 in the range
of about 1 mm; a thickness for lower spacer layer 506 in the range
of about 10 .mu.m; a thickness for waveguide bridge element 202 in
the range of about 10-20 .mu.m; a thickness for upper spacer layer
508 in the range of about 10 .mu.m; and a thickness for upper
substrate layer 504 in the range of about 1 mm. However, these
dimensions may vary significantly depending on the type of material
used.
[0033] For purposes of describing fabrication methods of an
integrated cross-connect device in accordance with a preferred
embodiment, FIG. 5 illustrates the lower substrate 502 as layer
"A", the lower spacer 506 as layer "B," the waveguide bridging
element 202 as layer "C", the upper spacer 508 as layer "D", and
the upper substrate 504 as layer "E". It is to be appreciated that
the each of these elements 502, 506, 202, 508, and 504 may comprise
multiple sub-layers or lateral sub-patterns of material (e.g., the
lower substrate layer 502 has waveguides and electrical plates
formed within it). However, for purposes of the device fabrication
steps described infra, it is simply required that any components of
a given layer material A, B, C, D, or E have common etching
characteristics with respect to a common set of etchants. Also, it
is required that some of the layer materials be responsive to
particular etchants to which other layer materials are not
responsive. A person skilled in the art would be readily able,
without undue experimentation, to determine proper sets of layer
compositions and chemical etchants to achieve the functionalities
described herein.
[0034] FIGS. 6-11 illustrate steps for fabricating a cross-connect
device in accordance with a preferred embodiment in which a wafer
bonding step is used. FIGS. 12-15 illustrate steps for fabricating
a cross-connect device in accordance with an alternative preferred
embodiment in which only a single-growth process is used.
[0035] FIG. 6 illustrates the results of a first step of
fabricating a cross-connect device according to the preferred
embodiment in which a wafer bonding step is used. In particular, a
top view and side view corresponding to a region of interest, or
cell, near the location of the K-K' cross-section of FIG. 2 is
illustrated. A simple three-layer structure is formed having no
patterns or cross-sections therein, comprising simply the layer "C"
as the top layer, the layer "B" as the middle layer, and the layer
"A" as the bottom layer. In FIG. 6, this is denoted by the string
"CBA" across the entire cell.
[0036] FIG. 7 illustrates the formation of a mask layer "M" over a
selected portion of the cell that defines the outline of the
waveguide bridging element. The masked region is denoted "MCBA" in
the top view of FIG. 7. The non-masked area just has the notation
"CBA." As indicated in the cross-section view of FIG. 7, the mask
"M" only covers a portion of the device along the line K-K'.
[0037] Following the mask application, the device is etched with an
etchant to which the C material, but not the B material, is
responsive. This etching step is preferably anisotropic so that the
lateral shape of the waveguide bridging element is maintained at
layer C. The result is illustrated in FIG. 8.
[0038] Next, an isotropic etching step is applied, using an etchant
that etches material B but does not etch material C or material A.
The result is illustrated in FIG. 9. As indicated in FIG. 9, the
elongated nature of the arc connector portion and neck of the
waveguide bridging element allows all of the B material under these
sections to be etched away, leaving only an air void in its place.
Notationally, the presence of a void in place of etched-away
material is shown as a "0." Thus, the string "C0A" over the arc
connector portion indicates there is the C material at the top
layer, air/void in the middle layer, and A material in the bottom
layer.
[0039] FIG. 10 illustrates the result of a standard formation of
the D and E layers. The D-E structure of FIG. 10 is formed on a
separate wafer than the A-B-C structure of FIG. 9. The two wafers
are then bonded together using wafer bonding techniques known in
the art, the resulting completed structure being illustrated in
FIG. 11.
[0040] FIG. 12 illustrates the results of a first step for
fabricating a cross-connect device according to the alternative
preferred embodiment in which only a single-growth process is used
and wafer bonding is avoided. In particular, a solid wafer is
formed comprising all of the required materials already deposited
in the proper place, except that material B is placed where air
gaps are supposed to be. In this embodiment, the lower spacer layer
(the second layer from the bottom) of the end result will comprise
layer C material instead of layer B material.
[0041] FIG. 13 illustrates the results of a second step in which a
mask covering particular portions of the device is applied. In
particular, a spacer area (MEDCCA) is masked along with the area of
the upper waveguide element as it exists on the E layer (MEBBBA,
MEBCBA, MEBBBA). Following the mask application, the device is
etched with an etchant to which E material, but not B material, is
responsive. This etching step is preferably anisotropic so that the
lateral shape of the waveguide is maintained at layer E. The result
is illustrated in FIG. 14.
[0042] Finally, the device of FIG. 14 is etched with an etchant to
which B material, but not A, C, D, or E material, is responsive.
This etching step should be isotropic so that proper undercuts are
achieved to form the desired air gaps. The result is illustrated in
FIG. 15. It is to be appreciated that the material "E" in the upper
left comer of the cross-sectional view of FIG. 15 is not actually
suspended in space. Rather, it is integral with the "E" layer of
the neighboring cell to the left of the cell illustrated in FIG.
15. In particular, it is integral with the portion of the
neighboring cell to the left corresponding to the "EDCCA" portion
of the cell illustrated in FIG. 15, and therefore is secured in
position. Likewise, the material "C" that appears to be hanging in
the cross-sectional view of FIG. 15 is actually integral with the
arc connector portion of the cell, so it too is secured in
position.
[0043] In fabricating a cross-connect device in accordance with the
preferred embodiments, one or more of the following references may
be of further assistance. Each of the following references is
incorporated by reference herein: U.S. Pat. Nos. 5,091,983
(Lukosz); 4,974,923 (Colak et. al.); 5,905,573 (Stallard et. al.);
5,506,919 (Roberts et. al.); 5,548,668 (Schaffner et. al.); and
4,471,474 (Fields); (a) Neukermans, A. et. al., "MEMS Technology
for Optical Networking Applications," IEEE Communications Magazine
(January 2001), pp. 62-69; (b) Lee, S. et. al., "Free-Space
Fiber-Optic Switches Based on MEMS Vertical Torsion Mirrors," J.
Lightwave Tech., Vol. 17, No. 1 (January 1999), pp. 7-13; (c)
Madsen, C. et. al., "A Tunable Dispersion Compensating MEMS
All-Pass Filter," IEEE Photonics Technology Letters, Vol. 12, No. 6
(June 2000), pp. 651-653; (d) Quevy, E. et. al., "Realization and
Actuation of Continuous-Membrane By an Array of 3D Self-Assembling
Micro-Mirrors For Adaptive Optics," Proceedings of the 14.sup.th
IEEE International Conference on Micro Electro Mechanical Systems
(2001), pp. 329-332; (e) Storment, C. et. al., "Flexible,
Dry-Released Process for Aluminum Electrostatic Actuators," J.
Microelectromechanical Systems, Vol. 3, No. 3 (September 1994), pp.
90-96; (f) Toshiyoshi, H. et. al., "Design and Analysis of
Micromechanical Tunable Interferometers for WDM Free-Space Optical
Interconnection," J. Lightwave Tech., Vol. 17, No. 1 (January
1999), pp. 19-25; (g) Lau, K.Y., "MEM's the Word for Optical Beam
Manipulation: Building Microelectromechanical-Based Optical Systems
on a Silicon Chip," Circuits & Devices (July 1997), pp. 11-18;
(h) Chen, J. et. al., "Optical Filters from Photonic Band Gap Air
Bridges," J. Lightwave Tech., Vol. 14, No. 11 (November 1996), pp.
2575-2580; (i) Muller, R. et. al., "Surface-Micromachined
Microoptical Elements and Systems," Proceedings of the IEEE, Vol.
86, No. 8 (August 1998), pp. 1705-1720; (j) Kwang, W. et. al., "A
New Flip-Chip Bonding Technique Using Micromachined Conductive
Polymer Bumps," IEEE Transactions on Advanced Packaging, Vol. 22,
No. 4 (November 1999), pp. 586-591; (k) Hammadi, S. et. al.,
"Air-Bridged Gate MESFET: A New Structure to Reduce Wave
Propagation Effects in High-Frequency Transistors," IEEE
Transactions on Microwave Theory and Techniques, Vol. 47, No. 6
(June 1999), pp. 890-899; (1) Ohta, A. et. al., "A 12-ps-Resolution
Digital Variable-Delay Macro Cell on GaAs 100 K-Gates Gate Array
Using a Meshed Air Bridge Structure," IEEE Journal of Solid-State
Circuits," Vol. 34, No. 1 (January 1999), pp. 33-41; (m)
Carts-Powell, Y., "MEMS Cantilever Controls a Guided-Wave Optical
Switch," WDM Solutions (January 2001), p. 9; and (n) Duvall, C.,
"VCSELs May Make Metro Networks Dynamic," WDM Solutions (November
2000), pp. 35-38.
[0044] Thus, according to one feature of the preferred embodiments,
an optical waveguide cross-connect device implemented in integrated
circuit form is provided, the cross-connect device comprising a
micromechanically-driven element that couples one of a plurality of
input waveguides to one of a plurality of output waveguides
responsive to an electrical signal.
[0045] According to another feature of the preferred embodiments,
such a cross-connect device is provided in which distinct bridging
waveguide elements, separate from the input and output waveguides
themselves, are used to bridge the input and output waveguides
together responsive to electrical control signals. In one preferred
embodiment, actual contact is made between the bridge waveguide
element and the input/output waveguides when in the ON state. In
another preferred embodiment, no contact is made but the bridge
waveguide is close enough to form resonant couplings with the
input/output waveguides when in the ON state.
[0046] According to another feature of the preferred embodiments,
the input and output waveguides are formed on first and second
layers of an integrated circuit structure, while the bridge element
is formed in a third layer of the integrated circuit structure
lying between the first and second layers.
[0047] According to another feature of the preferred embodiments,
an angle formed between the input array of waveguides and the
output array of waveguides is greater than 90 degrees, for
increasing the coupling efficiency. In still another preferred
embodiment, this angle exceeds 120 degrees.
[0048] According to another feature of the preferred embodiments,
the optical cross-connect device 100 is modified into a
cross-connect module such that a modular, expandable optical
cross-connect system may be established. According to this
embodiment, a single backplane may be provided that holds a
variable number of optical cross-connect modules, whereby a simple
M.times.N cross-connect system comprising a single cross-connect
module may be expanded by adding three identical cross-connect
switches to create a 2M.times.2N cross-connect system. To extend
the example, an additional five (5) identical cross-connect modules
may subsequently be added to form a 3M.times.3N cross-connect
system, and so on. In this manner, a single architecture and
component set may be used to form a wide variety of switch sizes.
By way of example and not by way of limitation, the individual
cross-connect modules may be 100.times.100 in size, such that an
original system of only 100.times.100 may later be expanded to form
a switch as large as 1000.times.1000 by the addition of ninety-nine
(99) additional modules.
[0049] FIGS. 16 and 17 illustrate conceptual diagrams of a modular
optical cross-connect system 1600 in accordance with a preferred
embodiment. For clarity of disclosure, only a 3.times.3 system is
illustrated (i.e., a system expandable from M.times.N to
3M.times.3N), although the number module slots may be substantially
larger. FIG. 16 shows the optical cross-connect system 1600 in
which only a single cross-connect module is inserted to create an
M.times.N optical cross-connect. FIG. 17 shows the optical
cross-connect system 1600 in which four (4) cross-connect modules
are inserted to create a 2M.times.2N optical cross-connect.
[0050] With reference to FIG. 16, optical cross-connect system 1600
comprises a backplane 1602 comprising a large array 1604 of
electrical surface contacts, an input coupling array 1606, and an
output optical coupling array 1608. A single optical cross-connect
module 1610 is provided having an internal switching fabric similar
to the integrated optical cross-connect embodiments described
supra. However, optical cross-connect module 1610 comprises input
waveguides that run all the way from an input edge (bottom edge in
FIG. 16) to the opposite end (top edge in FIG. 16), with the input
waveguide cross-sections appearing at both ends. Preferably, the
edge surfaces are extremely flat (e.g., surface variations less
than 0.4 .mu.m). The edge surfaces are adapted and configured such
that if a second cross-connect module is placed directly above a
first cross-connect module and precisely aligned therewith (e.g.
causing any gap therebetween to be not greater than 0.8 .mu.m at
any point along the border), an optical signal traveling along an
input waveguide in the first cross-connect module will exit that
waveguide, traverse the gap, enter into a corresponding input
waveguide in the second cross-connect module, and continue
propagating in the second cross-connect module. It has been found
that when the gap between the cross-connect modules is sufficiently
small (e.g., less than 0.8 .mu.m), any losses due to beam spreading
in the gap, and/or reflections back into the first cross-connect
module, are tolerable.
[0051] Similar to its input waveguides, the optical cross-connect
module 1610 further comprises output waveguides that run all the
way from the output edge (right edge in FIG. 16) to the opposite
edge (left edge in FIG. 16), these edges also being extremely flat.
These edges are also configured and dimensioned such that, if a
third cross-connect module is placed directly adjacent to the
output edge of the first cross-connect module (the right edge in
FIG. 16), an optical signal traveling along an output waveguide in
the first cross-connect module will exit that waveguide, traverse
the gap, enter into a corresponding output waveguide in the third
cross-connect module, and continue propagating in the third
cross-connect module. Because all four edges of the cross-connect
module 1610 are involved in light transfer in and out of the chip,
the electrical contacts of the cross-connect module 1610 are placed
on its bottom, whereby surface-mounting contact with the electrical
surface contacts 1604 may be achieved.
[0052] Input coupling array 1606 comprises optical elements (e.g.,
waveguide elements, micro-lenses, etc.) known in the art, and is
configured to be capable of receiving up to "3M" optical signals
from optical fibers, planar optical devices, or other input devices
and providing those light signals to the respective inputs of the
mounted cross-connect modules. Output coupling array 1608 is
likewise configured to be capable of receiving optical signals from
the output edges of the cross-connect modules and providing those
signals to external optical devices. Additionally, as illustrated
in FIGS. 16-17, output coupling array 1608 is designed to be
slidable in the direction of the output waveguides (i.e., the
horizontal direction in FIGS. 16 and 17) to mechanically
accommodate additional optical cross-connect modules as they are
added.
[0053] FIG. 17 shows the optical cross-connect system 1600 of FIG.
16, with additional cross-connect modules 1712, 1714, and 1716
placed on the backplane 1602 to form a 2M.times.2M switching array.
Advantageously, the cross-connect modules 1712, 1714, and 1716 may
each be identical to the initial cross-connect module 1610. Thus,
by providing a single backplane apparatus and one or more
integrated optical cross-connects in accordance with the preferred
embodiments, an expandable and modular optical switching
architecture is achieved.
[0054] In an alternative preferred embodiment, the output coupling
array 1608 may be fixably attached to the backplane 1602 instead of
being slidable. In this preferred embodiment, a plurality of
"dummy" extension chips are provided, comprising simple straight
waveguide arrays corresponding to the output waveguides of the
switching modules. In the event that fewer than the maximum number
of cross-connect chips are used, the extension chips are inserted
between the output edges of the cross-connect modules and the
output coupling array 1608.
[0055] Advantageously, the preferred embodiments described supra
may be adapted for use in a one-to-many "broadcast" type mode. For
this mode, the cross-connect elements within each cross-connect
module are made tunable, such that only a portion of the signal
energy in the input waveguide is extracted onto the output
waveguide at a given cross-connect. This allows the remaining
signal energy to be distributed among one or more additional output
waveguides. Likewise, a many-to-one switching fabric may also be
readily realized by the preferred embodiments described supra.
[0056] In another preferred embodiment, each cross-connect module
is equipped with beam size converters or lenses similar to the
coupling arrays 1608 and 1610. Consequently, each cross-connect
module is ready to receive and output collimated and spatially
extending beams. This beam conversion may significantly relax the
alignment precision and geometric precision when integrating many
modules onto a big backplane.
[0057] According to an additional preferred embodiment, with
particular application to the modular cross-connect system of FIGS.
16-17 in which signal loss may be problematic, a self-amplified
integrated cross-connect module is provided. The self-amplified
integrated cross-connect module is similar to the cross-connect
modules described supra except that the input waveguides, output
waveguides, and/or the arc connector portions of the waveguide
bridging elements are doped with erbium or other amplifying
element. Optical pumping signals may be introduced into each
waveguide sufficient to create an amplifying effect as the optical
signals pass through. According to one preferred embodiment, only a
single optical pumping input needs to be provided for an entire
cross-connect module, with the internal geometry being arranged
such that the pumping light is reflected and/or scattered around
the inside of the cross-connect module to bathe all waveguide
elements in the pumping light. This may be achieved at least in
part by providing mirrored surfaces along the top and bottom layers
of the cross-connect module, such that the pumping light is
reflected throughout a cavity formed by the mirrors.
[0058] Whereas many alterations and modifications of the present
invention will no doubt become apparent to a person of ordinary
skill in the art after having read the foregoing description, it is
to be understood that the particular embodiments illustrated and
described by way of illustration are in no way intended to be
considered limiting. By way of example, it is to be appreciated
that a person skilled in the art would be readily able to adapt the
methods and structures of the preferred embodiments to optical
cross-connects in which the incoming control signals are optical in
form, rather than electronic in form. This may be achieved, for
example, by using material that changes shape or size responsive to
optical signals in lieu of the electrostatically-driven elements,
or alternatively by providing an optical-to-electrical converter
having an output that drives the electrostatic control plates.
Therefore, reference to the details of the preferred embodiments
are not intended to limit their scope, which is limited only by the
scope of the claims set forth below.
* * * * *